Tom Wehmeier, partner and head of research at Atomico, and responsible for the annual State of European Tech report, and Gené Teare, Crunchbase senior data journalist, discuss the role of data in improving diversity in the tech and venture landscape. Taking the call for “more data” among many investors seriously, we go into the details of what we already know and how data can indeed drive change.
Interviewed January 2021
Why Data Matters for D&I in Tech
Johannes Lenhard (JL): What role do you think data plays in moving the needle on D&I in technology?
Gené Teare (GT): Crunchbase was one of the organizations that led on diversity data. From 2013 onwards, there was an active discussion around female representation on engineering teams. In early 2015, we recognized that with a global dataset of companies and founders in Crunchbase, we could provide analysis on funding to female founders, and how that has changed over time. In March 2015 we began to add gender to the data. We analyzed personal pronouns to help with data analysis, getting to a very comprehensive dataset within about two months, and wrote our first report. That first report showed a big shift over that time frame. We found that the percent of companies that had raised a first funding with at least one female founder had increased from around 9% in 2009 to 18% by 2014. Since then, we have continued to report how those numbers have shifted.
JL: How important do you think this tracking is? What do we get from it in a secondary and tertiary process, as the tracking in and of itself doesn’t change the needle, right?
GT: For those concerned about funding to female founders, without knowing exactly what the numbers are, you could over-project, under-project, or misunderstand the trends. The numbers themselves do not shift the needle, but understanding the numbers does anchor the discussion. It also can lead people to act because they say, “Okay, this is worse than I thought it was,” or “This isn’t moving quickly enough, we need to do more.” We have seen the data propel individuals to go out there and do something because they realize the numbers are not where they would like them to be.
Unlock expert knowledge.
Learn in depth. Get instant, lifetime access to the entire book. Plus online resources and future updates.
Tom Wehmeier (TW): I agree. I think data assists in helping to understand the scope of the issue and the potential for solving it. If we are not familiar with the scope of the challenges that we face, we won’t understand what levers we need to pull in order to shift the needle in the right way. I think data in general, whether it’s around D&I or anything else, is such a vital component of telling stories and doing so in a way that is objective. In doing the work to source the data, then using that to drive a data-driven narrative is a really powerful mechanism to raise awareness to engage people and close those knowledge gaps. When measured over time, data enables us to track progress and hold different stakeholders accountable. The last piece is that it can be incredibly empowering. It becomes easier to speak up about when you have data to support and validate your experience. All of these things, when taken together, shape how data plays a role in helping various different individuals or organizations to try to move the needle in the right direction.
How the State of European Tech Report Has Centered D&I
Erika Brodnock (EB): Tom, you oversee the State of European Tech report yearly, covering funding and developments in the industry, as well as a large survey of more subtle attitudes among VCs and founders—including many questions on DEI. Why is the State of European Tech report such an important piece of work at Atomico?
TW: I think it’s a small contribution to the discussion. I look back to the first report we delivered back in 2015, and I think a very limited amount of data was focused around gender and diversity at the founder and operator level. Over subsequent years, we’ve grown the level of focus in the report. In 2018, we really doubled down the level of data, specifically pulling out and creating a chapter around diversity and inclusion. We also used that as the lead narrative into all of the media briefings following publication.
We are fortunate that over the years the report has become a high-profile piece of research that reaches a large audience of influential people—investors, founders, policymakers, LPs, and others. We have always been aware that it is a powerful tool and platform to influence the industry agenda. Take 2020 as an example, hundreds of thousands of people have engaged with the report;* 9,000 people attended the launch event. We rarely measure social reach, but beyond the launch itself, there was a huge amount of engagement too.
Given the reach and potential influence of our reports, we felt a sense of responsibility to put diversity and inclusion at the front and center of what we do. Some of the things we do as part of the research process gives us a means by which we can gather information and data about D&I that wouldn’t otherwise be possible to collect at the scale that we can. Our survey, for example, features 5,000 voices from within the European tech industry. For multiple years now we have used that survey as a mechanism to collect interesting data around issues such as discrimination within the industry, or the extent to which companies that operate in the European tech industry are implementing different policies to drive positive progress around different aspects of D&I. Furthermore, we have some nice leverage in that when we have conversations with data partners that support us on the report, we can both ask them and push them to think through what data sets they might be able to share with their audience.
An incredibly important aspect is that it is a free report. It is open to anyone and each data point is exportable. The raw data, the charts, everything. This has been very deliberate, because we want it to be used as a resource that can be reused and recycled to create a compounding reach. It is a resource for anyone who wants to advocate for change. It has been great to see just how much people have taken that on, including the media. There are more than 500 media articles that we have tracked over the years that have a specific focus on D&I that reference back to some of the data in our reports.
However, we are still far from having anything close to all the answers, there are still mega hurdles in terms of collecting information about the state of what is happening. It is sobering to think that there might be people out there who think that what we do goes beyond what others do. Yet internally, we are there thinking this is only 5% of what it should be. It is amazing that we have people like Gené, with Crunchbase, who have for many, many years now been looking to address this too. As the diversity discussion has evolved beyond gender, Crunchbase has been at the forefront looking at ways to paint a clearer picture with data on different aspects of diversity.
Collecting the Data We Need to Drive Change
JL: Gené, what do you think is the most important data that we can collect to change the landscape? Give some examples of things that you have already found and dive into some of the results.
GT: At Crunchbase, we track companies as soon as they first raise funding, and then they start showing up on our reports. What is critical is that Crunchbase is an open platform that founders can access, to add themselves and their early funding rounds and become a visible part of the ecosystem.
In 2020, we added race and ethnicity to the dataset for the US market as part of our Diversity Spotlight data. People had been asking us over the years to look at the data with a race and ethnic lens. We had previously thought this was too difficult to address as race and ethnicity is sensitive data, and the meaning of race varies in different parts of the world. People would need to self-report at scale for it to be meaningful. We took the decision that, as a global database, it was too challenging.
In 2020, specifically in response to Black Lives Matter, we decided to add race and ethnicity to the data. We started with the US market, as it is a leading funding market. The process we went through for this data was interesting; we partnered extensively to get feedback and collect the data with organizations active in the Black and Latinx community, as well as investors. We worked with organizations such as the Black Founder List, Black VC, Stanford Latino Entrepreneurship Initiative, and All Raise, and the venture community.
After updating companies that were started by Black or Latinx founders, our findings confirmed what many were saying. “Within the US the funding to Black and Latinx founders is low, we think it’s 1%, we think it’s 2%”—everyone was collating their own datasets and coming up with those findings. We thought, given that Crunchbase has all this data, when we build it out we can do a comprehensive job and possibly show improvement on these stark statistics, whilst recognizing that we will miss some companies that will be added over time. What we found is that funding to Black founders was around or just under 1% over the last five years. When we looked at Latinx founders, that number was just under 2%. However, if you look at those two communities, Latinx make up 18.5% of the US population based on the US Census Bureau, and the African American community came in at 13.4%.
What’s meaningful for us is that having worked with these partners, they don’t need to recreate the dataset. They can take the Crunchbase data to then inform their own reporting on these issues.
Expanding to Europe
EB: I work with Extend Ventures, and we have published similar reporting for the UK. We are really interested in looking at the numbers across Europe. Do you have plans for that?
GT: Yes, we would like to explore expanding our diversity data to Europe. For Diversity Spotlight we tag companies, not individuals, with “founded” or “led” based on race or ethnicity. By “led,” we are focused on the CEO role at a company.
How Research Translates into Programs
EB: Tom, how has Atomico used the research findings to enhance its thesis? What are the particular programs and investments that you’re making now?
TW: Atomico’s founding thesis back in 2006 was, “Great companies can come from anywhere.” And that is what we have learned. We have spent more time understanding the industry issues around D&I, that “anywhere” really means “by anyone in any location.” The more we have spent time understanding this, the more it has reinforced and strengthened our conviction in that initial thesis, specifically, our resolve to back underrepresented talent in tech. It has been really influential in helping rethink every aspect of what we do, in terms of how we invest, also how we think about building our team as a fund.
To highlight a few specific initiatives that have directly come off the back of the work that we’ve done around D&I: we run an angel program and are launching our third cohort of angels. The first cohort was run in 2018. The primary motivation was to play a role in building a more diverse universe of early-stage angel investors in Europe. Assessing the numbers, we can see that, significantly greater than 90% of all angels and significantly greater than 95% of all investments involve white male angels, and so right from the outset of that program, D&I was embedded. Across the first two cohorts, 60% of the angels were women and 25% came from an underrepresented ethnicity. One of the core beliefs we held was, if we brought a more diverse set of people into the program, they would go back and invest into a more diverse set of founders. That group of 30 or so angels have made 150 investments across two years; 64% of those investments had a founder from either an underrepresented gender or ethnicity. It has been rewarding to see how research translated into a thesis, that translated into a direct initiative, that is now translating into amazing results in terms of helping to drive allocation of funding to a more diverse set of founders.
Ultimately, this is a long-term initiative for us, because our hope is very much that those companies that get backed with angel investments will mature and develop to become the next the companies that we back from our core fund. Also, from that program, the deals that we do are a direct reflection of the network that we maintain. It is therefore critical for us to ensure that we have a diverse network ourselves; inherent in getting to a more diverse network, is building a more diverse internal team. This has been reinforcing how we think about driving diversity into our investment team, as a direct means of driving more diversity into the investments that we make.
Some of the research we did highlighted the industry’s dependency on warm introductions and the barriers to entry and blocks to network diversity this creates. Understanding that warm intros form a barrier to more diverse ways of capital allocation led us to roll out our Access Atomico program. We scaled a program that runs permanently and allocates time from every member of our investment team to be available for people to apply to directly.
More broadly, research has shown there is a lack of diversity at the GP layer in Europe at the early stages, which are formative moments for any young company gaining its first institutional check. That is problematic for us as a Series A investor, because clearly the companies that flow through to us are a reflection of the companies that get backed by the funds that invest before us. We created an initiative off the back of some of those insights to basically run a small fund-of-funds program. We are now making LP commitments to early-stage emerging fund managers with a clear focus on investing in underrepresented founders.
Turning the Tide of Warm Introductions
EB: Tom, how are you selecting the angels? If the angels are selected from a warm introduction style network, surely that kind of perpetuates the problem, so is it open?
TW: We start by going through our network to find candidates to go to our shortlist. We then leverage the network of the angels already on the program to make recommendations too. Though I would say that they are sought-out recommendations from our network, we are thoughtful in who we go and ask to supply those introductions to ensure that that field of people is diverse. There is also then the opportunity to openly apply. We have also been proactive in a data-driven way, to identify a potential diverse pool of candidates that we have reached out to ourselves, because we thought that they might represent interesting candidates for the program.
The search for potential angels has actually highlighted another problem that is not just at the founder level. Essentially, when you look at the founder level, then you look at the senior executive level across venture-backed startups, there is very high lack of diversity at that specific level. That matters given what we know about the flywheel of how those people go on to become the next generation of founders, and in having that profile. If you cannot break that lack of diversity at the senior executive level too, it just becomes cyclical and really hard to break.
Facing the Next Set of Challenges
JL: Where do you think we need more data for the future? Crunchbase started collecting data on women eight years ago, and on ethnicity relatively recently. What do we not yet know enough about, and what is another frontier that we can help push with data?
GT: I spoke with BBG Ventures founders Susan Lyne and Nisha Dua, who focus on investing in consumer technology companies with a female founder. One of their findings is that their incoming deal flow in 2020 during COVID-19 for female founders increased. We spoke with other funds focused on female founders and confirmed this finding. That contrasts with our most recent report, demonstrating funding in 2020 to female-only founded companies was down, tracking closer to 2% than the previous year’s 3%. That is concerning. We also cite a McKinsey report, which mentions 30% of working mothers may leave the workforce due to COVID-19. The concern at this point is: are we going to see stagnation or a step backwards for women in the workplace?
Tom, your point about the lack of diversity in the senior leadership level is an indicator that the next set of founders might not be as diverse as we expect. From my perspective, looking at the data in 2014, 18% of companies with an initial funding had at least one female founder. This includes female-only and female/male co-founded companies. The numbers have shifted a couple of percentage points to about 20% of companies in recent years. We haven’t seen the same leap in the last five years that we did in the previous five years.
However, absolute funding has gone up considerably. In 2016, total global funding was around $170B. In the last three years, funding has almost doubled, closer to $300B per year on a global basis. Based on percentages, the pace of change has been slower. From an absolute number of companies founded by a woman and from an absolute number of dollars, however, the counts and amounts have grown compared to the previous five years.*
Tom made a point about networks being incredibly important. All Raise, which supports female founders here in the US, were concerned during COVID-19 that video conferencing doesn’t quite replace the traditional networking function. The fact that people can’t meet at events serendipitously, they feel, has impacted women because networks are incredibly important in this business.
TW: I agree. The things that would be most impactful is just building a more comprehensive coverage of some of the research and initiatives that we are already undertaking, whether that is building out coverage across regions, for gender or ethnicity. Also, being able to comprehensively map and understand the full lifecycle of a company. Not just in terms of their investments, but also be able to comprehensively look at things like exits and returns and IPO pathways. I think we need to be able to connect all of those different dots through the lens of gender or through the lens of ethnicity, and to start putting even more data on the table around the relative performance and returns from different founder groups. We are probably still at a point where we will still find amazing datasets that emerge from this level of research, but often, it’s snapshots in time. I would love to have continuous longitudinal time series data across some of these areas that offers deeper coverage—then we would be in a really powerful position to tell the real story of what is happening.
But I would just say that every incremental 10% of improvement in the data that does exist is to be celebrated. It is not about trying to get to 100%. Every incremental report helps and is a valuable contribution to how we will eventually be able to create lasting and systemic change.